Optimizing the Bit-flipping Method for Decoding Low-density Parity-check Codes in Wireless Networks by Using the Artificial Spider Algorithm

Abstract

In this paper, the performance of Low-Density Parity-Check (LDPC) codes is improved, which leads to reduce the complexity of hard-decision Bit-Flipping (BF) decoding by utilizing the Artificial Spider Algorithm (ASA). The ASA is used to solve the optimization problem of decoding thresholds. Two decoding thresholds are used to flip multiple bits in each round of iteration to reduce the probability of errors and accelerate decoding convergence speed while improving decoding performance. These errors occur every time the bits are flipped. Then, the BF algorithm with a low-complexity optimizer only requires real number operations before iteration and logical operations in each iteration. The ASA is better than the optimized decoding scheme that uses the Particle Swarm Optimization (PSO) algorithm. The proposed scheme can improve the performance of wireless network applications with good proficiency and results. Simulation results show that the ASAbased algorithm for solving highly nonlinear unconstrained problems exhibits fast decoding convergence speed and excellent decoding performance. Thus, it is suitable for applications in broadband wireless networks.

Keywords

• a low-density parity-check (ldpc)
• hard-decision bit-flipping (bf)
• particle swarm optimization (pso)
• artificial spider algorithm (asa)